Arc vs InfluxDB
Arc is 5x faster than InfluxDB 3 Core on ingestion and 1.6x faster on analytical queries, with Telegraf support out of the box, portable Parquet storage, and standard SQL instead of deprecated Flux.
Query Benchmark (vs DataFusion)
InfluxDB 3 is not on ClickBench, but its query engine (Apache DataFusion) is. Arc is benchmarked against DataFusion directly: same Parquet format, same hardware, same 43 queries. Verify on benchmark.clickhouse.com →
Combined Score (lower is better)
| System | Machine | Score |
|---|---|---|
| Arc | c8g.metal-48xl | ×1.04 |
| Arc | c7a.metal-48xl | ×1.13 |
| Arc | c6a.4xlarge | ×1.92 |
| DataFusion (Parquet, single) | c6a.4xlarge | ×2.99 |
Cold Run (lower is better)
| System | Machine | Score |
|---|---|---|
| Arc | c8g.metal-48xl | ×1.15 |
| Arc | c6a.4xlarge | ×1.21 |
| Arc | c7a.metal-48xl | ×1.21 |
| DataFusion (Parquet, single) | c6a.4xlarge | ×1.79 |
Hot Run (lower is better)
| System | Machine | Score |
|---|---|---|
| Arc | c8g.metal-48xl | ×1.01 |
| Arc | c7a.metal-48xl | ×1.15 |
| Arc | c6a.4xlarge | ×2.80 |
| DataFusion (Parquet, single) | c6a.4xlarge | ×5.10 |
Ingestion Benchmark
Sustained 60-second ingestion load. MacBook Pro M3 Pro Max, same record schema, Line Protocol over HTTP for InfluxDB.
| System | Throughput | Batch size | p50 latency |
|---|---|---|---|
| Arc | 17.3M rec/s | 1,000 rows | 4.4ms |
| InfluxDB 3 Core | 3.45M rec/s | 100,000 rows | 259ms |
| InfluxDB 3 Enterprise | 1.71M rec/s | 100,000 rows | 590ms |
Arc achieves 5x higher throughput than Core and 10x vs Enterprise, with 100x smaller batches and single-digit millisecond latency.
Why Arc Is Different: Under the Hood
Both Arc and InfluxDB 3 store data as Parquet and use a vectorized query engine. The differences are in data accessibility, write handling, compaction, and what you get without an Enterprise license.
Storage Format
Self-describing paths vs. a metadata catalog
Both systems store data as Apache Parquet. Arc organizes files in a self-describing time-partitioned path (db/measurement/YYYY/MM/DD/HH/) readable by any Parquet-compatible tool without going through Arc. InfluxDB 3 tracks file locations in a Postgres-compatible metadata catalog: the files are standard Parquet, but discovering them without the catalog requires knowing its internal layout. InfluxDB 3 Core also has no background compactor, so small Parquet files accumulate over time and query performance degrades. Compaction is an Enterprise-only feature.
Query Engine
DuckDB vs. DataFusion: 1.6x gap on the same Parquet
Arc embeds DuckDB (C++, vectorized OLAP). InfluxDB 3 uses Apache DataFusion (Rust, vectorized). Both are modern vectorized engines. On ClickBench with the same Parquet format and hardware, Arc scores ×1.04 to ×1.92 vs DataFusion's ×2.99, a 1.6x gap. Arc also applies SQL rewrites before execution: regex calls become string functions, time bucketing uses epoch arithmetic. InfluxDB 3 Core restricts queries to a 72-hour window; querying historical data beyond that requires Enterprise.
Ingestion Protocol
Optional WAL vs. mandatory WAL
Both systems accept InfluxDB Line Protocol, so existing Telegraf pipelines work with both without changes. Arc also accepts MessagePack binary columnar batches (18M+ rec/s) and bulk Parquet import. The throughput gap comes from WAL handling: Arc's WAL is optional (disable for maximum throughput, enable for crash safety). InfluxDB 3's WAL is mandatory, which adds latency on every write: 3.45M rec/s at 100K-row batches vs Arc's 17.3M rec/s at 1,000-row batches.
Deployment Model
One binary vs. binary + Postgres catalog
Arc is a single Go binary with no required external dependencies. Clustering uses embedded Raft with no separate coordinator. InfluxDB 3 Core runs as a single binary for development, but production deployments require a Postgres-compatible database for the metadata catalog. That catalog tracks table schemas, partition locations, and compaction state, and needs to be backed up, replicated, and sized separately. Combined with the missing compactor in Core, a production-grade InfluxDB 3 deployment effectively requires Enterprise.
Feature Comparison
| Feature | Arc | InfluxDB 3 |
|---|---|---|
| Standard SQL (no proprietary language) | ✓ | ✓ |
| Portable Parquet storage | ✓ | ✓ |
| Open source | ✓ | ✓ |
| InfluxDB Line Protocol ingestion | ✓ | ✓ |
| Edge / single-binary deployment | ✓ | ✗ |
| WAL can be disabled for max throughput | ✓ | ✗ |
| Retention policies | ✓ | ✓ |
Frequently Asked Questions
Can I migrate from InfluxDB to Arc without changing my Telegraf setup?
Yes. Arc supports InfluxDB Line Protocol natively. Point your Telegraf output plugin at Arc's write endpoint and change the URL. No collector changes, no schema migration, no downtime.
Why is InfluxDB not on ClickBench?
InfluxDB 3 has not been submitted to ClickBench. Its underlying query engine, Apache DataFusion, is on ClickBench. We benchmarked Arc against DataFusion directly: same Parquet format, same hardware, same 43 queries. DataFusion is the closest available proxy for InfluxDB 3's query performance.
Does Arc support Flux?
No, and you don't need it. InfluxData deprecated Flux in InfluxDB 3. Arc uses DuckDB SQL, a PostgreSQL-compatible dialect with full analytical features: window functions, CTEs, joins, and subqueries. No proprietary query language required.
How does WAL affect ingestion performance?
InfluxDB 3 always writes through a Write-Ahead Log, which adds latency on every write and limits throughput. Arc's WAL is optional: disable it for maximum throughput (18M+ rec/s), enable it for zero data loss on crash. You choose the tradeoff.
Pricing
Start free with open source. Scale with enterprise features when you need them.
Open Source
- 18M records/sec ingestion
- Full SQL query engine (DuckDB)
- Parquet storage (S3, GCS, local)
- Docker and Kubernetes ready
- Community support (Discord)
Arc Cloud
Managed hosting. No infrastructure. Free 30-day trial.
- Deploy in 30 seconds
- Dedicated physical servers
- Daily backups to S3
- Arc Enterprise included
- No credit card required
Enterprise
Starting price for up to 8 cores. Clustering, RBAC, and dedicated support.
- Everything in Open Source
- Horizontal clustering and HA
- Role-based access control (RBAC)
- Tiered storage and auto-aggregation
- Dedicated support and SLAs
Enterprise Features
Clustering
Horizontal scaling with automatic data distribution. Query routing and load balancing across nodes.
Security
Fine-grained RBAC with database and table-level permissions. LDAP/SAML integration available.
Data Management
Automated retention policies, continuous queries for aggregation, and tiered storage for cost optimization.
Ready to handle billion-record workloads?
Deploy Arc in minutes. Own your data in Parquet. Use for analytics, observability, AI, IoT, or data warehousing.